1. Field of the Invention
The invention relates to a reflex triode for use in producing ultra-high-current (&gt;10.sup.5 A), ultra-high-power (.gtoreq.10.sup.11 W) ion beams.
2. Description of the Prior Art
An ion beam producing reflex triode (shown in FIG. 1) is a device consisting of an electron emitting cathode and an anode which is a thin film that is semi-transparent to electrons. The anode is at high positive potential relative to the cathode during an applied electrical pulse. Electrons emitted from the cathode are accelerated toward the anode, pass through it and form a virtual cathode. The anode is located between the two cathodes. As a consequence of the energy dissipated at the anode and elastic scattering which reduces the electron axial velocity, the electrons perform damped axial oscillations. The ions are extracted out of the plasma formed from the anode film by the oscillating electrons and by surface flashover. Some ions are accelerated toward the real cathode and some are accelerated toward the virtual cathode. The ions that are accelerated toward the virtual cathode pass through it and form a drifting beam that is current-and space-charge-neutralized.
Humphries, Lee and Sudan (Applied Physics Letter 25, p. 20, 1974) were the first to report results on the production of ion beams using reflex trioxides. In their first reported attempt, the anode of the reflex triode was constructed of parallel copper wires coated with varnish. Because the varnished wires provided meager plasma, only weak ion beams were obtained (i.e., about 250 A (one side) ion current having energy of about 100 keV). In their second attempt, Humphries, Lee and Sudan replaced the parallel copper wires with nylon filaments. The difficulty with this approach, as with the previous one, is that the plasma is produced only near the plastic filaments. As a consequence of this fact, the ion-current density is low and the beam is non-uniform or badly divergent. These difficulties with non-uniformity are further exacerbated by operating the device in high axial magnetic fields which constrain the transverse excursions of the electrons. A more important limitation of this configuration is the absence of inelastic energy loss and elastic scattering by the oscillating electrons at the anode. This loss of energy and the reduction of axial velocity which results from elastic scattering is important because after the electron passes through the anode it cannot pass through as large a potential difference toward the virtual or real cathode as prior to that transmited through the anode. The result of the axial velocity loss is a piling up of electron space charge near the anode which results in an enhanced ion-beam production.
High voltage (megavolt) operation of an ion-beam producing reflex triode was attempted by S. Humphries, Sudan and Condit (Applied Physics Letter 26, p. 667, 1975). Using an anode made of either aluminum foil, mylar film, or nylon mesh, only weak ion beams were produced. Although the 1.8 MV, 30.OMEGA. pulse generator theoretically could produce 60 kA cathode current with a proton current as large as 30 kA, the maximum ion current produced by the reflex triode appeared to be either 2.5 kA of protons or 5 kA of aluminum ions. In this experiment the triode presented the enormous load inductance of 1.4 .mu.H to the pulse generator. Most of the energy delivered to the reflex triode by the pulse generator went into inductive magnetic field generation and not into the ion and electron flows of the triode which constitute the resistive load of the device.
The previous experiments above point out four of the most serious deficiencies of prior art reflex triodes. The first deficiency is with the design of the anode and its inability to produce a uniformly adequate anode plasma which is necessary for a uniform beam. The second problem deals with the high inductance of the reflex triode and the resultant loss in electrical efficiency owing to energy being spent in inductive magnetic fields. The third deficiency is in the anode design which does not lead to axial velocity reduction with each pass while still permitting a large number of electron transits through the anode; thus, the ratio of ion current (one way) to "real-cathode" current is much less than 0.5. The fourth problem is that the anodes and cathodes of prior-art triodes have not been configured to provide a uniform axial electric field and well-defined virtual cathode; thus, they produced beams with large divergence.
The first and third deficiencies noted above were partially resolved by the use of an anode consisting of a polyethylene film interwoven among parallel thin copper wires (0.75 mm diameter). The use of this anode is described in J. Golden, C. A. Kapetanakos, Proc. of the 1st International Topical Conference on Electron Beam Research and Technology, Vol. I, Albq. N.M. Nov. 3-5, 1975, p. 635, and in C. A. Kapetanakos, J. Golden, W. M. Black, Phys. Review Lett. 37, 1236, 1976. This anode had two serious drawbacks. First the thin wires were not capable of withstanding the energy deposition of the reflexing electron for 1 MV, 2.times.10.sup.10 W, 0.2 kA/ cm.sup.2 ion beam operating levels. Second, the wires whose purpose was to help define the anode equipotential surface and turn on the cathode electron emission had the deleterious effect of limiting the number of electron transits through the anode.
Several aspects of the first three deficiencies were also resolved in experiments by D. S. Prono, J. M. Creedon, I. Smith, and N. Bergstrom (J. Applied Phys. vol. 46, p. 3310, 1975) and by D. S. Prono, J. W. Shearer, and R. J. Briggs (Phy. Rev. Lett. v. 37, p. 21, 1976). In their studies a large negative voltage pulse was applied to the cathode. The anode was a thin metallic or polymer film which was connected to a grounded conducting vacuum chamber. Although this prior art did have a low inductance design, the configuration had the serious limitation that the chamber was at the same potential at which the ions originate in the anode plasma so that ions would decelerate when approaching near the chamber walls thus making beam extraction difficult. Moreover, for successful operation, it was necessary to use a tenuous neutral background gas filling to obtain an adequate virtual cathode. In practice the device operated at very low impedances and with inefficient coupling to the pulse generator.